Add file read ahead through AsyncDataCache #3389
Closed
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
✅ Deploy Preview for meta-velox canceled.
|
|
@oerling has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
Yuhta
reviewed
Dec 1, 2022
| auto nextQuantum = position_ - offsetInQuantum + loadQuantum_; | ||
| auto prefetchThreshold = loadQuantum_ * prefetchPct_ / 100; | ||
| if (!prefetchStarted_ && offsetInQuantum + *size > prefetchThreshold && | ||
| position_ - offsetInQuantum + loadQuantum_ < region_.length) { |
There was a problem hiding this comment.
nextQuantum < region_.length?
| // Remove the loads that were done. There can be done loads if the same | ||
| // CachedBufferedInput has multiple cycles of enqueues and loads. | ||
| for (int32_t i = doneIndices.size() - 1; i >= 0; --i) { | ||
| allCoalescedLoads_.erase(allCoalescedLoads_.begin() + doneIndices[i]); |
There was a problem hiding this comment.
This should be outside the looping over allCoalescedLoads_.
Also this is more efficient if allCoalescedLoads_ is large:
std::swap(allCoalescedLoads_[doneIndices[i]], allCoalescedLoads_.back());
allCoalescedLoads_.pop_back();|
@oerling has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
1 similar comment
|
@oerling has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
|
@oerling has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
c1c8bb0 to
fb50810
Compare
|
@oerling has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
Yuhta
approved these changes
Dec 1, 2022
oerling
pushed a commit
to oerling/velox-1
that referenced
this pull request
Dec 1, 2022
Summary: adds a code sample and supporting functions for using CachedBufferedInput for smart prefetching of sequential files. A SeekableInputStream is registered for each file of interest. Each time the read proceeds past a given fraction of the current load quantum, the load of the next quantum is scheduled. The load prefetches into AsyncDataCache and will be found there when moving past the end of thecurrent quantum. Prefetch will fail silently if there is no memory or if over half the cache is taken by not yet accessed prefetched data. Adds a special StreamIdentifier to denote expected sequential access. The default for no StreamIdentifier in enqueue is preloading the whole enqueued range. Adds a mode where each visited cache entry is made evictable immediately after unpinning. This allows not polluting the cache with one time large sequential accesses. Adds a test that simulates a multifile merge. Each thread has 100 files and consumes a chunk of each in turn. Pull Request resolved: facebookincubator#3389 Reviewed By: Yuhta Differential Revision: D41642368 Pulled By: oerling fbshipit-source-id: 3272aca53cc3beb7b8a5489418802d552b92441e
|
This pull request was exported from Phabricator. Differential Revision: D41642368 |
Summary: adds a code sample and supporting functions for using CachedBufferedInput for smart prefetching of sequential files. A SeekableInputStream is registered for each file of interest. Each time the read proceeds past a given fraction of the current load quantum, the load of the next quantum is scheduled. The load prefetches into AsyncDataCache and will be found there when moving past the end of thecurrent quantum. Prefetch will fail silently if there is no memory or if over half the cache is taken by not yet accessed prefetched data. Adds a special StreamIdentifier to denote expected sequential access. The default for no StreamIdentifier in enqueue is preloading the whole enqueued range. Adds a mode where each visited cache entry is made evictable immediately after unpinning. This allows not polluting the cache with one time large sequential accesses. Adds a test that simulates a multifile merge. Each thread has 100 files and consumes a chunk of each in turn. Pull Request resolved: facebookincubator#3389 Reviewed By: Yuhta Differential Revision: D41642368 Pulled By: oerling fbshipit-source-id: 86adf8ea400aed76f2681c8f291f3d619ea4f492
|
This pull request was exported from Phabricator. Differential Revision: D41642368 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
adds a code sample and supporting functions for using CachedBufferedInput for smart prefetching of sequential files. A SeekableInputStream is registered for each file of interest. Each time the read proceeds past a given fraction of the current load quantum, the load of the next quantum is scheduled. The load prefetches into AsyncDataCache and will be found there when moving past the end of thecurrent quantum. Prefetch will fail silently if there is no memory or if over half the cache is taken by not yet accessed prefetched data.
Adds a special StreamIdentifier to denote expected sequential access. The default for no StreamIdentifier in enqueue is preloading the whole enqueued range.
Adds a mode where each visited cache entry is made evictable immediately after unpinning. This allows not polluting the cache with one time large sequential accesses.
Adds a test that simulates a multifile merge. Each thread has 100 files and consumes a chunk of each in turn.